TWI796275B - Method for disease risk assessment - Google Patents
Method for disease risk assessment Download PDFInfo
- Publication number
- TWI796275B TWI796275B TW111131783A TW111131783A TWI796275B TW I796275 B TWI796275 B TW I796275B TW 111131783 A TW111131783 A TW 111131783A TW 111131783 A TW111131783 A TW 111131783A TW I796275 B TWI796275 B TW I796275B
- Authority
- TW
- Taiwan
- Prior art keywords
- subject
- disease
- risk assessment
- physiological information
- disease risk
- Prior art date
Links
Images
Landscapes
- Measuring And Recording Apparatus For Diagnosis (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
Abstract
Description
本發明係關於一種風險評估相關技術,尤指一種疾病風險評估方法。 The invention relates to a technology related to risk assessment, in particular to a disease risk assessment method.
現有之疾病風險評估系統是針對受測者的健康檢查報告或受測者所填寫的問卷,而分析受測者的靜態生理數據、生活習慣以及家族健康史而取得疾病風險評估結果。 The existing disease risk assessment system analyzes the static physiological data, living habits, and family health history of the subject to obtain the disease risk assessment result based on the subject's health examination report or the questionnaire filled out by the subject.
然而,一般根據健檢中心或醫療單位提供的健康檢查所取得的檢查數據報告,只能反應受測者當下及當次的健康狀況,以在臨床上進行侷限性的病理判斷,例如以肝功能、腎功能異常作為進一步檢查的依據,且相關數據的常模範圍往往沒有差異性,不分男女老少都參考相同的數值,例如血壓正常範圍皆為120/80mmHg,因此不易作為未來健康狀況風險的評估。 However, the examination data report obtained from the health examination provided by the health examination center or medical unit can only reflect the current and current health status of the subject, so as to make limited pathological judgments in clinical practice, such as liver function , Abnormal renal function as the basis for further examination, and the normative range of relevant data is often no difference, regardless of gender, age and age, all refer to the same value, for example, the normal range of blood pressure is 120/80mmHg, so it is not easy to be used as an indicator of future health risk Evaluate.
此外,很多的疾病早期徵狀會展現在身體動作與反應的改變上,單就受測者的健康檢查數據報告與問卷並無法獲悉受測者神經系統的異常,當然沒有辦法進行相關疾病風險的評估,也就無法提早改善健康狀況與預防疾病發生。 In addition, many early symptoms of the disease will be manifested in changes in body movements and reactions. The testee’s health check data report and questionnaire alone cannot know the abnormalities of the subject’s nervous system, and of course there is no way to assess the risk of related diseases Therefore, it is impossible to improve the health status and prevent the occurrence of diseases in advance.
本發明主要目的在於透過定期與長時間的健康趨勢評估,能夠評估受測者未來罹患特定疾病的可能性,以提醒受測者應提早改善健康狀況以預防或延緩疾病發生。尤其是指一種透過定期的評估受測者執行特定動作的動態特徵,來評估受測者罹患具動作障礙特徵的疾病的可能性,以提醒受測者應提早改善健康狀況以預防或延緩疾病發生。 The main purpose of the present invention is to assess the possibility of a subject suffering from a specific disease in the future through regular and long-term health trend assessment, so as to remind the subject to improve their health status in advance to prevent or delay the occurrence of the disease. In particular, it refers to a method to assess the possibility of a subject suffering from a disease characterized by movement disorders by regularly evaluating the dynamic characteristics of the subject's execution of a specific action, so as to remind the subject to improve their health status early to prevent or delay the occurrence of the disease .
為達上述目的,本發明提供一種疾病風險評估方法,其包含一擷取步驟、一前處理步驟以及一判斷步驟;於擷取步驟中,利用一擷取模組取得一受測者之一醫療紀錄、受測者所量測的一靜態生理資訊與對應不同動作的一動態生理資訊,靜態生理資訊係躺透過一靜態生理檢測儀器量測受測者於動作靜止時的生理特徵而取得,動態生理資訊係透過對受測者傳達複數動作指令,而由一動態生理檢測儀器量測受測者執行各動作指令的生理特徵而取得;於前處理步驟中,利用一終端裝置整合醫療紀錄、靜態生理資訊以及動態生理資訊以產生一當前數據;於判斷步驟中,將當前數據輸入至一預測模型進行演算,以產生受測者之一疾病風險評估結果,疾病風險評估結果為一疾病種類、對應疾病種類之發病可能性以及發病預估時間。 To achieve the above purpose, the present invention provides a disease risk assessment method, which includes an extraction step, a preprocessing step, and a judgment step; in the extraction step, an extraction module is used to obtain a medical Record, a static physiological information measured by the subject and a dynamic physiological information corresponding to different actions. The static physiological information is obtained by measuring the physiological characteristics of the subject when the action is still by lying on a static physiological detection instrument. Dynamic Physiological information is obtained by transmitting multiple action instructions to the subject, and a dynamic physiological detection instrument measures the physiological characteristics of the subject's execution of each action instruction; in the pre-processing step, a terminal device is used to integrate medical records, static Physiological information and dynamic physiological information to generate a current data; in the judgment step, the current data is input into a prediction model for calculation to generate a disease risk assessment result of the subject, the disease risk assessment result is a disease type, corresponding The possibility of onset of the disease type and the estimated time of onset.
其中,預測模型是透過輸入已知罹患疾病受測者的醫療紀錄、靜態生理資訊與動態生理資訊進行訓練與建立,尤其是同時包含確診罹患某種疾病前期與確診時期的醫療紀錄、靜態生理資訊與動態生理資訊。而一位尚未罹患疾病的受測者於確診罹患某種疾病後,則其過去與當下的資料將成為預測模型更新參數的新的訓練案例。 Among them, the prediction model is trained and established by inputting the medical records, static physiological information and dynamic physiological information of the subjects known to suffer from diseases, especially the medical records and static physiological information of the pre-diagnosed and diagnosed period. and dynamic physiological information. After a subject who has not yet suffered from a disease is diagnosed with a certain disease, his past and current data will become a new training case for updating the parameters of the prediction model.
藉此,本發明疾病風險評估方法透過擷取受測者的醫療紀錄,以 及量測受測者的靜態生理資訊與動態生理資訊而整合產生的當前數據,並將當前數據輸入至預測模型,能夠全面性的分析受測者過去、當下的健康狀況,以及受測者的活動力與身體控制能力,而取得受測者未來之疾病風險評估結果,進而瞭解受測者未來健康狀態的變化趨勢,以利於受測者提早進行健康控管而改善身體狀況,降低罹病風險且預防或延緩疾病發生。 Thus, the disease risk assessment method of the present invention extracts the medical records of the subjects to And the current data generated by integrating the static physiological information and dynamic physiological information of the subject, and inputting the current data into the prediction model, can comprehensively analyze the past and current health status of the subject, as well as the subject's Activity and body control ability, so as to obtain the results of the future disease risk assessment of the subject, and then understand the change trend of the subject's future health status, so as to facilitate the early control of the subject's health to improve the physical condition, reduce the risk of disease and Prevent or delay the onset of disease.
100:評估系統 100: Evaluation System
10:終端裝置 10: Terminal device
11:擷取模組 11: Capture module
12:處理模組 12: Processing module
13:預測模型 13: Prediction Model
14:資料庫 14: Database
20:靜態生理檢測儀器 20: Static physiological testing equipment
30:動態生理檢測儀器 30: Dynamic physiological testing equipment
200:醫療就醫系統 200: Medical treatment system
S1:擷取步驟 S1: Extraction step
S2:前處理步驟 S2: Preprocessing steps
S3:記錄步驟 S3: Record step
S4:判斷步驟 S4: Judgment step
圖1係本發明實施例之方塊連結示意圖。 Fig. 1 is a schematic diagram of block connections of an embodiment of the present invention.
圖2係本發明實施例之疾病風險評估方法流程圖。 Fig. 2 is a flowchart of a disease risk assessment method according to an embodiment of the present invention.
圖3係本發明實施例之疾病風險評估之系統流程圖。 Fig. 3 is a system flow chart of disease risk assessment in the embodiment of the present invention.
為便於說明本發明於上述發明內容一欄中所表示的中心思想,茲以具體實施例表達。實施例中各種不同物件係按適於說明之比例、尺寸、變形量或位移量而描繪,而非按實際元件的比例予以繪製,合先敘明。 In order to illustrate the central idea of the present invention expressed in the column of the above-mentioned summary of the invention, it is expressed in specific embodiments. Various objects in the embodiments are drawn according to proportions, sizes, deformations or displacements suitable for illustration, rather than drawn according to the proportions of actual components, which are described first.
請參閱圖1至圖3所示,本發明提供一種疾病風險評估方法,本發明之疾病風險評估方法係透過一評估系統100而執行。
Please refer to FIG. 1 to FIG. 3 , the present invention provides a disease risk assessment method, and the disease risk assessment method of the present invention is implemented through an
評估系統100包含一終端裝置10、一靜態生理檢測儀器20以及一動態生理檢測儀器30,如圖1所示,終端裝置10與靜態生理檢測儀器20、動態生理檢測儀器30以及一醫療就醫系統200耦接。
The
終端裝置10能夠取得一受測者之一醫療紀錄,靜態生理檢測儀器
20能夠量測受測者於動作靜止時的生理特徵而取得一靜態生理資訊,動態生理檢測儀器30能夠量測受測者於執行對應不同動作時的生理特徵而取得一動態生理資訊。
The
於本實施例中,終端裝置10可為各種雲端運算裝置、個人電腦、筆記型電腦、智慧型行動裝置或平板電腦等裝置;靜態生理檢測儀器20可為血糖計、血中藥物濃度感測器、體溫計、血壓計、體重計、身高計、握力計、視力檢測計、聽力檢測計、握力計、皮尺(胸圍、腹圍、腿圍量測)、體脂計、腦波機、心電圖機或血氧計等,亦可以結合複數上述儀器作為靜態生理檢測儀器20;動態生理檢測儀器30可為體溫計、血壓計、血氧計、心電圖機、腦波機、肌電感測器、足底壓力感測器、慣性量測系統、動作擷取系統或攝影裝置等,亦可以結合複數上述儀器作為動態生理檢測儀器30。
In this embodiment, the
於本實施例中,醫療紀錄包括受測者歷史就診之藥品藥理治療分類碼與處方箋,醫療紀錄更包括受測者之診斷日期、診斷病名以及診斷病名之國際疾病分類碼(ICD-10)。 In this embodiment, the medical records include the drug pharmacological treatment classification codes and prescriptions of the subject's history of visits, and the medical records also include the subject's diagnosis date, diagnosis disease name, and International Classification of Diseases (ICD-10) of the diagnosis disease name .
於本實施例中,靜態生理資訊選自受測者之血液檢測值、體溫、身高、體重、身高、體脂率、雙手握力、胸圍、腹圍、小腿圍、血壓、血氧、視力、聽力、血糖、靜態心電圖、靜態腦波圖其中之一者或其組合。 In this embodiment, the static physiological information is selected from the blood test value, body temperature, height, weight, height, body fat percentage, grip strength of both hands, chest circumference, abdominal circumference, calf circumference, blood pressure, blood oxygen, vision, Hearing, blood sugar, static electrocardiogram, static electroencephalogram, or a combination thereof.
於本實施例中,動態生理資訊包括複數代表不同生理特徵的時域訊號,所述時域訊號選自連續體溫偵測、連續血壓偵測、連續血氧偵測、心電訊號、腦波電訊號、肌肉電訊號、足底壓力訊號、身體軀幹動作慣性訊號、動作擷取訊號以及影像訊號其中之一者或其組合。 In this embodiment, the dynamic physiological information includes a plurality of time-domain signals representing different physiological characteristics, and the time-domain signals are selected from continuous body temperature detection, continuous blood pressure detection, continuous blood oxygen detection, ECG signals, and brain wave signals. One or a combination of signals, electromyography signals, plantar pressure signals, trunk movement inertial signals, motion capture signals, and video signals.
終端裝置10具有相互耦接之一擷取模組11、一處理模組12、一預
測模型13以及一資料庫14,擷取模組11網路連結至醫療就醫系統200而取得受測者之醫療紀錄,擷取模組11通訊連接靜態生理檢測儀器20以及動態生理檢測儀器30而取得靜態生理資訊以及動態生理資訊,處理模組12整合受測者的醫療紀錄、靜態生理資訊以及動態生理資訊形成當前數據,並將當前數據與過去儲存於資料庫14的同一受測者的過去數據輸入至預測模型13,而分析得所述受測者之一疾病風險評估結果,其中,疾病風險評估結果為一疾病種類、對應該疾病種類之發病可能性以及發病預估時間。當前數據將儲存到資料庫14中,供受測者下一次進行檢測時,預測模型13進行分析時做為過去數據或未來預測模型13的更新訓練資料。
The
上述內容,為說明本發明之評估系統100的一具體實施例,以下進一步說明本發明之疾病風險評估方法,如圖2至圖3所示,圖2為疾病風險評估方法流程圖,圖3為疾病風險評估之系統流程圖,疾病風險評估方法包含一擷取步驟S1、一前處理步驟S2、一記錄步驟S3以及一判斷步驟S4。
The above content is to illustrate a specific embodiment of the
於擷取步驟S1中,利用擷取模組11連結至醫療就醫系統200而取得受測者最近一次之醫療紀錄,利用靜態生理檢測儀器20量測受測者於動作靜止時(如站立不動、坐著或平躺)的生理特徵而取得靜態生理資訊,並透過對受測者傳達複數動作指令,而利用動態生理檢測儀器30量測受測者執行各動作指令的生理特徵而取得動態生理資訊。
In the retrieval step S1, the
於本實施例中,對受測者所傳達之動作指令係選自於張眼單腳站立、張眼雙腳站立、張眼前後腳站立、張眼路腳站立、閉眼單腳站立、閉眼雙腳站立、閉眼前後腳站立、閉眼路腳站立、起身、坐下、蹲下、直線行走、轉彎及登階等,或者多個上述指令的組合。 In this example, the action instructions conveyed to the subjects were selected from standing on one foot with eyes open, standing on two feet with eyes open, standing on one foot with eyes open, standing on one foot with eyes open, standing on one foot with eyes closed, and standing on two feet with eyes closed. Standing, standing with eyes closed, standing with feet closed, standing with closed eyes and feet, getting up, sitting down, squatting down, walking in a straight line, turning and climbing steps, etc., or a combination of multiple above commands.
於前處理步驟S2中,終端裝置10之處理模組12整合受測者之醫療紀錄、靜態生理資訊以及動態生理資訊以產生一當前數據,當前數據係藉由處理模組12對醫療紀錄、靜態生理資訊以及動態生理資訊進行同步處理以及特徵分析所生成。
In the pre-processing step S2, the
較特別的是,縱使是單一個動作指令,例如直線行走,當聽到往前行走的指令到跨出第一步即可視為一個事件,開始走動後的步態週期中,左右腳各別的離地與騰空也是不同事件,若沒有精確標記各個事件的時間點,一個高齡長輩的受試者,與早期的帕金森氏症患者,二位受試者從聽到指令到完成第一步的時間可能一樣,但是由於帕金森氏症患者會有凍凝步態(Freezing of gait)症狀,導致他們起步所花的時間不同,因此除了必須要能標示出起步階段與左右步伐階段以計算時間的差異外,更進一步的需要將「起步」與「左右步伐」區分為不同事件,並標示在腦波電訊號、肌肉電訊號、足底壓力訊號與身體動作影像的時間訊號上,才能在分析時針對各動態生理資訊依據不同事件進行精準的關聯性與差異性分析。 What is more special is that even if it is a single action command, such as walking in a straight line, it can be regarded as an event when hearing the command to walk forward and taking the first step. In the gait cycle after starting to walk, the left and right feet separate Landing and vacating are also different events. If the time points of each event are not accurately marked, an elderly subject and an early Parkinson's disease patient, the time from hearing the instruction to completing the first step of the two subjects may be The same, but because Parkinson's patients have symptoms of freezing gait (Freezing of gait), they take different times to start, so in addition to being able to mark the difference between the start phase and the left and right gait phases to calculate the time Further, it is necessary to distinguish "start" and "left and right steps" into different events, and mark them on the time signals of brain wave signals, muscle signals, plantar pressure signals and body motion images, so as to target each event during analysis. Dynamic physiological information conducts precise correlation and difference analysis based on different events.
進一步說明,動態生理資訊的事件同步是在進行動態生理資訊量測階段,當特定事件發生時,即時的將特定事件以編號或編碼為事件旗標,並將事件旗標標記於每一個動態生理訊號量測設備所量測的時域訊號資料中事件旗標標記產生的時間點是一個指令、動作或特殊事件的開始或結束的當下,而使各該時域訊號在基於事件進行特徵分析處理時,事件產生的時間點一致。標記的方式可以透過人工手動按壓事件標示器、預先設計好的導引設備於導引時同步觸發事件標示器或由某一個動態生理檢測儀器30觸發事件標示器,由事件標示器發出事件旗標標記訊號,供其它的動態生理檢測儀器30接收並標註於各自紀錄
的動態生理訊號內,此外也可以由處理模組12先分析某一個動態生理檢測儀器30的訊號做為事件旗標標記。
Further explanation, the event synchronization of dynamic physiological information is during the dynamic physiological information measurement phase. When a specific event occurs, the specific event is numbered or coded as an event flag in real time, and the event flag is marked on each dynamic physiological information. The time point when the event flag mark is generated in the time domain signal data measured by the signal measurement equipment is the beginning or end of an instruction, action or special event, so that each time domain signal is analyzed and processed based on the event , the time point of event generation is the same. The way of marking can be manually pressing the event marker, the pre-designed guiding device triggers the event marker synchronously during guidance, or triggers the event marker by a certain dynamic
而動態生理資訊的特徵分析係進一步按受測者執行動作指令的順序,將各生理資訊的時域訊號依據各別事件的事件旗標分割為複數同步區塊,該區塊可以由二個事件旗標所定義,或是以單一事件旗標為起點往前或往後一固定秒數,亦或是以單一事件旗標為中間點,往前與往後一固定秒數,再依據各同步區塊之間的關聯性、時間差、週期變化、協調性以及重心轉移趨勢以生成複數特徵指標。其中,所述特徵指標係關於受測者於各個事件階段自主動作時之肌力、重心轉移能力、平衡能力、神經感覺敏銳度、神經協調性與反應速度;於本實施例中,複數同步區塊分別包括一指令接收區塊、一動作意念區塊、一起步區塊、一加速區塊、一維持區塊、一減速區塊、一停止區塊以及一意念結束區塊。 The characteristic analysis of dynamic physiological information further divides the time-domain signal of each physiological information into multiple synchronous blocks according to the sequence of the subject's execution of action instructions according to the event flag of each event. This block can be composed of two events. Defined by the flag, either start from a single event flag and go forward or backward by a fixed number of seconds, or use a single event flag as the middle point, go forward and backward by a fixed number of seconds, and then synchronize according to each The correlation, time difference, periodic change, coordination and center of gravity shift trend between blocks are used to generate complex characteristic indicators. Wherein, the characteristic index is related to the muscular strength, center of gravity transfer ability, balance ability, nerve sensory acuity, nerve coordination and reaction speed of the subject during voluntary actions at each event stage; The blocks respectively include a command receiving block, an action idea block, a start block, an acceleration block, a maintenance block, a deceleration block, a stop block and an idea end block.
需特別說明的是,自主動作是來自於腦部意念區先有動作的意念,動作意念會引發受測者腦部運動皮質區運作,而觸發控制肌肉的動作區去操控相對應的肌肉群組產生肢體動作以及肌電反應,並透過感覺回饋區接收感覺神經系統所感受到的環境資訊,而整體肌肉群之運作效果極大部分會反應到足底之壓力分佈與重心變化。 It should be noted that the voluntary action comes from the idea of the previous action in the brain's idea area. The action idea will trigger the operation of the motor cortex area of the subject's brain, and trigger the action area of the control muscles to control the corresponding muscle group Generate body movements and myoelectric responses, and receive environmental information sensed by the sensory nervous system through the sensory feedback area, and the operation effect of the overall muscle group will largely reflect the pressure distribution and center of gravity changes on the soles of the feet.
因此,處理模組12可以結合動態生理資訊之肌肉電訊號與腦波電訊號,取得肌肉與腦部皮質層神經運動活性的關聯性與時間差,而生成關於各個事件階段自主動作的特徵指標。
Therefore, the
處理模組12亦可以由足底壓力分佈訊號取得不同步態週期訊號,同時比對肌肉電訊號,分析出肌肉群的出力與時序狀況以及足底壓力變換產生重心轉移趨勢,而生成關於重心的特徵指標。
The
以常見用於評估平衡感的Timed Up and Go(TUG)test為例,對受測者進行步態的同步處理以及特徵分析進行說明,假設對受測者所傳達的動作指令順序為從坐在椅子上開始,「起身」後「直線行走」到3公尺處的一個圓錐,然後「轉彎」,「直線行走」,回到原處接著「坐下」。藉由如攝影系統、肌電圖-腦電圖共調性設備、足底壓力感測器等動態生理檢測儀器30取得受測者進行TUG test的動態生理資訊,處理模組12首先以攝影機訊號為基礎,分析「起身」、「直線行走」、「轉彎」、「坐下」等,每一動作的時間點做為事件旗標,並將其標示於其他的動態生理資訊的訊號內,既可對整體動態生理資訊的各生理特徵之時域訊號進行同步處理,而使各時域訊號在「起身」、「直線行走」、「轉彎」以及「坐下」的各階段動作有可供訊號分析的一致時間點。
Taking the Timed Up and Go (TUG) test, which is commonly used to evaluate the sense of balance, as an example, the synchronization processing and feature analysis of the gait of the subject are explained, assuming that the sequence of action instructions conveyed by the subject is from sitting to Start on the chair, "get up", "walk straight" to a cone at 3 meters, then "turn", "walk straight", return to the original position and then "sit down". The dynamic physiological information of the subject undergoing the TUG test is obtained by dynamic
於本實施例中,處理模組12對「直線行走」進一步分析其時域訊號,並分割為指令接收區塊、動作意念區塊、起步區塊、加速區塊、維持區塊、減速區塊、停止區塊以及意念結束區塊。
In this embodiment, the
其中,處理模組12能夠將受測者起步區塊、加速區塊、維持區塊、減速區塊以及停止區塊的影像訊號重建為人體骨架,從視覺角度紀錄受測者的空間動作,將動作區再分為站立與擺動(Stance Phase & Swing Phase)二種相位,並結合肌電訊號而取得不同相位下肌肉群的出力與時序狀況,以生成關於肢體協調的特徵指標。
Among them, the
於本實施例中,處理模組12進一步結合受測者動作意念區塊、起步區塊、加速區塊、維持區塊、減速區塊、停止區塊以及意念結束區塊的肌電訊號與腦電訊號,而取得動作與腦部皮質層神經運動活性的關聯性,進而生成關於神經感覺敏銳度、神經協調性與反應速度的特徵指標。
In this embodiment, the
於本實施例中,處理模組12更進一步藉由起步區塊、加速區塊、維持區塊、減速區塊以及停止區塊的足底壓力訊號,取得受測者的足底重心轉移趨勢,以生成關於重心轉移與平衡能力的特徵指標。
In this embodiment, the
進一步說明,從受測者的自主動作的控制能力、指令下達與產生動作的延遲時間以及不同動作之間的協調程度,能夠分析受測者的衰老程度或健康狀況,而從受測者的活動力能夠分析受測者動作障礙的嚴重程度,以輔助後續疾病風險的判斷。因此,藉由特徵分析動態生理資訊而生成的特徵指標,能夠更深入瞭解受測者在身體自主動作的控制能力以及瞭解受測者的活動力,從而判斷受測者發生疾病的風險。 To further explain, from the control ability of the subject's voluntary actions, the delay time between giving instructions and producing actions, and the degree of coordination between different actions, it is possible to analyze the subject's aging degree or health status, and from the subject's activities It is able to analyze the severity of the subject's movement disorder to assist in the subsequent judgment of disease risk. Therefore, the feature index generated by feature analysis of dynamic physiological information can provide a deeper understanding of the subject's ability to control voluntary body movements and understand the subject's activity, thereby judging the subject's risk of disease.
於前處理步驟S2中,處理模組12更能夠自醫療紀錄中取得受測者歷史就診之藥品藥理治療分類碼與處方箋,以及診斷病名與其國際疾病分類碼,以從受測者用藥中判斷是否會影響受測者所量測的靜態生理資訊以及動態生理資訊。另一方面,透過取得就診紀錄中使用之藥品的國際疾病分類碼可以確定受測者所欲治療的問題種類,進一步可作為後續疾病風險評估的依據。
In the pre-processing step S2, the
於記錄步驟S3中,處理模組12能夠將受測者每一次經由前處理步驟S2所產生的當前數據,按照整合的日期儲存至資料庫14作為受測者之歷程紀錄。
In the recording step S3, the
於判斷步驟S4中,處理模組12將受測者之當前數據,以及受測者過去其他時間點的歷程紀錄輸入至預測模型13進行演算,利用受測者的歷程紀錄變化以及完整的歷程紀錄資料,以輔助預測模型13生成疾病風險評估結果,並避免使用單一時間點的數據而有可能發生誤判的問題。
In the judgment step S4, the
舉例來說,受測者今年為58歲,其自52歲開始每間隔三個月利用
評估系統100進行疾病風險評估,假設所述受測者在最近一次檢查的前兩週因感冒而持續服用感冒藥,而導致在最近一次的當前數據顯示其具有肝功能異常,由於受測者以往的歷程紀錄資料都沒有顯示其肝功能指標有異常,因此預測模型13能夠排除受測者在肝功能導致的相關疾病的評估而避免造成誤判。
For example, the subject is 58 years old this year, since he was 52 years old, he used
The
另外,當前數據輸入至預測模型13後,同時會依據預存於資料庫14對應不同疾病之一特徵權重,對所述當前數據進行權重分配與計算可能發病的疾病以及發病機率,其中,特徵權重包含分別對所述當前數據之醫療紀錄、靜態生理資訊以及動態生理資訊之複數特徵指標的權重參數。
In addition, after the current data is input into the
於本實施例中,特徵權重的分配原則為首先依不同的疾病而分配權重參數,接著再依受測者醫療紀錄中的用藥資料與處方箋再次增加或減少用藥資料所造成的相關權重參數分配,藉此降低藥物對受測者的影響所產生的干擾,以更為精準的進行特定疾病的風險評估。 In this embodiment, the distribution principle of feature weights is to first assign weight parameters according to different diseases, and then increase or decrease the distribution of related weight parameters caused by medication data and prescriptions in the subject’s medical records , so as to reduce the interference caused by the influence of drugs on the subjects, so as to carry out more accurate risk assessment of specific diseases.
舉例來說,透過目前技術水平得知對於罹患中風症狀高相關的因子為「高血壓、肥胖、膽固醇以及舒張壓」,便可以將上述因子對於判斷罹患中風的權重參數做適當的增加,而符合實際狀況;事實上,也有可能透過AI人工智能經過數據的收集,便可理解此一趨勢而由AI人工智能進行適當的參數模型調整。 For example, based on the current technical level, it is known that the factors highly correlated with the symptoms of stroke are "high blood pressure, obesity, cholesterol, and diastolic blood pressure", and the above factors can be appropriately increased for the weight parameters of the judgment of stroke, which meets the The actual situation; in fact, it is also possible to understand this trend through the collection of data through AI artificial intelligence and make appropriate parameter model adjustments by AI artificial intelligence.
於判斷步驟S4中,預測模型13能夠進一步將當前數據與預測模型13預設之案例比對而推算發病預估時間,以產生疾病風險評估結果。藉由比對當前數據之特徵指標與預設案例,能夠根據案例的大數據統計結果取得受測者與大群體的差異性,而自案例所建立的特徵層級推算受測者發病預估時間;於本實施例中,案例係與特定疾病以及特定執行動作相關,而根據人種、國別、性別、
年齡層的受測者進一步統計與分類所建立。舉例來說,以受測者走路時的步態特徵與關於中風的步態預設案例比對為例,假設受測者為亞洲66歲之男性,當預測模型13接收到受測者之當前數據後,會將所述當前數據之對應走路相關的特徵指標與對應於亞洲65歲至70歲男性的步態案例進行比對,而推算出其可能罹患中風的風險與時間。
In the judgment step S4, the
於本實施例中,預測模型13之建立方式係收集醫療紀錄中有明確疾病分類碼之複數受測者每次評估的動態生理資訊、靜態生理資訊與醫療紀錄為基礎,利用一機器學習演算法而建立。尤其是讓預測模型13找出各種疾病被明確診斷時間點之前的各種醫療紀錄、靜態生理資訊以及動態生理資訊的變化趨勢,一旦某一個尚未罹患疾病的新的受測者的相關數據變化依循著某種已知疾病發病的趨勢與時程變化,就可評估該受測者可能罹患該種疾病的機率與可能發病時間點。
In this embodiment, the establishment of the
關於預測模型13的更新,係透過持續的增加複數尚未罹患疾病的受測者,每一受測者每次經由擷取模組11連結至醫療就醫系統200而取得之醫療紀錄也將會偵測其疾病分類碼項目是否有更新,針對有新增或更新疾病分類號的受測者,其醫療紀錄、當前數據與自資料庫14取得過往之檢測歷程紀錄,都將用於預測模型13的驗證、重新訓練與更新,以生成持續演進之預測模型13。特別的是,受測者發生重大疾病後,可能無法再使用評估系統100進行評估,因此,資料庫14也會定時擷取所有受測者的醫療就醫紀錄,以判斷是否有受測者發生重大疾病,此類發生重大疾病的受測者資料也是預測模型13重要的驗證、重新訓練與參數更新的寶貴資料之一。
The update of the
進一步說明,預測模型13之訓練結合案例式推理(Case-Based
Reasoning)技術,藉由持續累積罹患疾病的受測者案例,利用該些有明確疾病標記受測者的生理變化以及自主動作能力變化趨勢,而增加預測模型13的訓練資料,而達到持續演進預測模型13,且逐步提高預測模型13準確度之效。
To further illustrate, the training of the
舉例來說,假設一年齡為66歲的亞洲女性受測者透過評估系統100進行帕金森氏症的評估,帕金森氏症為一種影響中樞神經系統的慢性神經退化疾病,帕金森氏症病發的前兆為運動功能衰退以及走路時步態異常。
For example, assume that a 66-year-old Asian female subject is assessed for Parkinson's disease through the
於本實施例中,所述受測者自63歲起每個月都有透過評估系統100記錄與評估。因此,於擷取步驟S1中,取得所述受測者其最新的醫療紀錄、當下的靜態生理資訊以及動態生理資訊;於前處理步驟S2中,處理模組12自資料庫14取得所述受測者過去的歷程紀錄,再整合所述受測者的靜態生理資訊以及動態生理資訊而生成當前數據;於記錄步驟S3中,處理模組12將所述受測者的當前數據記錄至資料庫14;於判斷步驟S4中,處理模組12將所述受測者的當前數據輸入至預測模型13進行演算,由於歷程紀錄包含所述受測者在63歲至66歲時每個月都有服用可能造成血壓升高藥物,並領有慢性處方籤,且在65歲至66歲的一年內,關於步態的特徵指標有每三個月減少2公分的情況,關於重心轉移能力的特徵指標有下降的趨勢,而執行動作指令期間血壓則有緩慢增加的趨勢,在經由預測模型13演算,可以得知受測者在未來九個月內可能罹患帕金森氏症,患病機率為80%。藉由疾病風險評估結果,所述受測者能夠提早透過服藥以及進行復健運動,而改善身體活動機能以及提升生活品質,以延緩帕金森氏症的發生。
In this embodiment, the subject has been recorded and evaluated through the
綜合上述,本發明之疾病風險評估方法透過擷取受測者的醫療紀錄,以及量測受測者的靜態生理資訊與動態生理資訊而整合產生的當前數據,能夠藉由處理模組12以及預測模型13全面性的分析受測者過去與當下的健康狀
況,以及受測者的活動力與身體控制能力,而取得受測者之疾病風險評估結果,進而瞭解受測者未來健康狀態的變化趨勢,以利於受測者提早進行健康控管而改善身體狀況,降低罹病風險且預防或延緩疾病發生。
To sum up the above, the disease risk assessment method of the present invention captures the subject’s medical records and measures the subject’s static physiological information and dynamic physiological information to integrate the current data, which can be processed by the
此外,藉由持續且大量的收集不同受測者的資料,尤其是有新的疾病被診斷出來的受測者資料,能夠不斷訓練以及驗證預測模型13而使其持續演進,強化預測模型13的疾病判斷能力,令其他尚未罹患疾病的受測者取得準確的罹患疾病可能性以及發生時間。
In addition, by continuously and massively collecting the data of different subjects, especially the data of subjects diagnosed with new diseases, the
以上所舉實施例僅用以說明本發明而已,非用以限制本發明之範圍。舉凡不違本發明精神所從事的種種修改或變化,俱屬本發明意欲保護之範疇。 The above-mentioned embodiments are only used to illustrate the present invention, and are not intended to limit the scope of the present invention. All modifications or changes that do not violate the spirit of the present invention belong to the intended protection category of the present invention.
10:終端裝置 10: Terminal device
11:擷取模組 11: Capture module
12:處理模組 12: Processing module
13:預測模型 13: Prediction Model
14:資料庫 14: Database
20:靜態生理檢測儀器 20: Static physiological testing equipment
30:動態生理檢測儀器 30: Dynamic physiological testing equipment
200:醫療就醫系統 200: Medical treatment system
Claims (10)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW111131783A TWI796275B (en) | 2022-08-24 | 2022-08-24 | Method for disease risk assessment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW111131783A TWI796275B (en) | 2022-08-24 | 2022-08-24 | Method for disease risk assessment |
Publications (2)
Publication Number | Publication Date |
---|---|
TWI796275B true TWI796275B (en) | 2023-03-11 |
TW202410067A TW202410067A (en) | 2024-03-01 |
Family
ID=86692423
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW111131783A TWI796275B (en) | 2022-08-24 | 2022-08-24 | Method for disease risk assessment |
Country Status (1)
Country | Link |
---|---|
TW (1) | TWI796275B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TWI845343B (en) * | 2023-06-13 | 2024-06-11 | 國泰人壽保險股份有限公司 | Disease probability prediction system and disease probability prediction method |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TW202143248A (en) * | 2020-05-13 | 2021-11-16 | 國家中山科學研究院 | System and method of biomedical data prediction risk including a data collection unit, a data processing unit and a judgment unit |
-
2022
- 2022-08-24 TW TW111131783A patent/TWI796275B/en active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TW202143248A (en) * | 2020-05-13 | 2021-11-16 | 國家中山科學研究院 | System and method of biomedical data prediction risk including a data collection unit, a data processing unit and a judgment unit |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TWI845343B (en) * | 2023-06-13 | 2024-06-11 | 國泰人壽保險股份有限公司 | Disease probability prediction system and disease probability prediction method |
Also Published As
Publication number | Publication date |
---|---|
TW202410067A (en) | 2024-03-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Rovini et al. | How wearable sensors can support Parkinson's disease diagnosis and treatment: a systematic review | |
Chen et al. | Toward pervasive gait analysis with wearable sensors: A systematic review | |
Vernon et al. | Quantifying individual components of the timed up and go using the kinect in people living with stroke | |
JP6914896B2 (en) | Health condition diagnosis system | |
EP3403572A1 (en) | Pulse wave diagnostic device and method of calculating and determining fatigue | |
CN113168895B (en) | Method, apparatus and program for assessing the correlation between the health of a health care zone and various preventive intervention actions | |
US11622729B1 (en) | Biomechanics abnormality identification | |
AU2010286471A1 (en) | Characterizing a physical capability by motion analysis | |
de Carvalho Bastone et al. | Validation of the human activity profile questionnaire as a measure of physical activity levels in older community-dwelling women | |
US20220330879A1 (en) | Systems and methods for monitoring the state of a disease using a biomarker, systems and methods for identifying a biomarker of interest for a disease | |
WO2020216816A1 (en) | Device for calculating, during one step or each successive step of the gait of a subject, the push-off p0 of the subject | |
Rao et al. | Accelerometers for objective evaluation of physical activity following spine surgery | |
US20230298760A1 (en) | Systems, devices, and methods for determining movement variability, illness and injury prediction and recovery readiness | |
Exley et al. | Predicting updrs motor symptoms in individuals with Parkinson's disease from force plates using machine learning | |
TWI796275B (en) | Method for disease risk assessment | |
Moreau et al. | Overview on wearable sensors for the management of Parkinson’s disease | |
Chen et al. | Sarcopenia Recognition System Combined with Electromyography and Gait Obtained by the Multiple Sensor Module and Deep Learning Algorithm. | |
CN112674760A (en) | Wearable sensor-based Parkinson upper limb movement detection method | |
US10966652B1 (en) | Method and system for quantifying movement disorder systems | |
Sant'Anna et al. | A wearable gait analysis system using inertial sensors part ii-evaluation in a clinical setting | |
US20240127958A1 (en) | Method for disease risk assessment | |
Tokmak et al. | Unveiling the relationships between seismocardiogram signals, physical activity types and metabolic equivalent of task scores | |
TW202004773A (en) | System for diagnosing cognitive function for providing fitness correction program and method thereof | |
Qiang et al. | Auxiliary Diagnosis for Parkinson’s Disease Using Multimodal Feature Analysis. | |
Shammas et al. | Detection of parameters to quantify neurobehavioral alteration in multiple sclerosis based on daily life physical activity and gait using ambulatory assessment |